“AI fluency and data competence are not the same thing,” says Andreas Bartsch, Head of Innovation and Services at PBT Group. “Someone can be very comfortable using an AI tool and still lack the experience needed to recognise when an answer looks technically sound but is actually wrong, incomplete, or unsuitable for the business context.”
For organisations building data capability, this changes the skills conversation. Teaching people how to use AI tools is increasingly important, but that training needs to sit alongside the foundations that enable a Data Specialist to test assumptions, challenge outputs, and understand the systems and data behind them.
The foundations have survived every technology shift
Bartsch has spent three decades working in the data industry and has seen successive generations of technology change how data teams work. The move from on-premises environments to cloud platforms, along with IoT, event-driven capabilities, and modern data architectures, required people to learn new technologies. The underlying disciplines remained relevant.
Experienced data engineers and analysts could often move between platforms because their knowledge was not tied to one product or tool.
“AI is another major change in the technology available to us, but it does not remove the need to understand data. The fundamentals give people something to reason from when the technology changes,” says Bartsch.
Those foundations include practical skills such as SQL and data modelling, together with an understanding of how data moves through pipelines and how reliability is tested. Numerical aptitude, logical reasoning, troubleshooting, and problem-solving are equally important because they help specialists recognise when an output needs to be questioned.
AI can generate a query. The specialist still needs to understand whether it is efficient, whether the joins make sense, whether the underlying assumptions are valid, and whether the result addresses the business problem.
Junior talent needs room to build judgement
The challenge is particularly relevant for younger data professionals. Many are comfortable experimenting with AI and can become productive with new tools very quickly. What takes longer to develop is the experience that helps somebody recognise when a plausible answer should not be trusted.
A junior engineer still needs to understand why a data model was designed in a certain way, what happens when assumptions inside a pipeline change, and how data quality affects the result.
Without that grounding, AI can create confidence before professional judgement has had time to develop. Bartsch says organisations should not respond by discouraging junior employees from using AI.
“The opportunity is to develop AI fluency alongside strong data foundations. Junior professionals should be encouraged to use these tools, but they also need enough understanding to interrogate the output and take responsibility for the work they put forward,” he says.
There is another side to the skills gap. Senior data professionals often bring the experience needed to challenge AI-generated work, while some may be more hesitant to adopt AI capabilities themselves. Bringing those strengths together gives teams an opportunity to develop younger specialists without losing the judgement built through experience.
AI training needs to include accountability
At PBT Group, AI-enablement extends beyond teaching consultants how to use available tools. Internal guidance emphasises protecting confidential information, understanding the business context, verifying AI-generated output, and taking ownership of work presented to a client or colleague.
“As soon as you put your name to an output, you own it. AI can help you get there faster, but the responsibility for checking the logic, understanding the context, and deciding whether the result is reliable remains with the person doing the work,” says Bartsch.
This places greater emphasis on how organisations develop data talent. AI literacy belongs in the toolkit, alongside the disciplines that allow specialists to assess quality, understand architecture and governance, and work through a problem when the obvious answer does not hold.
Judgement becomes more valuable
As AI makes some routine technical work easier and faster to produce, the value of a Data Specialist shifts further towards architecture, data integrity, governance, and business understanding.
Data professionals will increasingly be expected to know how to use AI and when to challenge it. For Bartsch, that is the standard organisations should build towards.
“The strongest Data Specialists will be people who are comfortable using AI and understand its output well enough to question it. Speed is useful, but only when it is supported by the competence needed to know whether the result can be trusted,” concludes Bartsch.